{
 "id": "AIR-2026-0341",
 "slug": "data-exfiltration-via-mcp-server-prompt-injection-targeting",
 "title": "Data Exfiltration via MCP Server Prompt Injection Targeting Cursor",
 "summary": "Researchers showed that a malicious Model Context Protocol server can deliver an indirect prompt injection to the Cursor IDE. This injection tricked the AI agent into executing shell commands to exfiltrate configuration files.",
 "type": "attack_on_ai",
 "categories": [
  "prompt_injection",
  "agent_hijack",
  "data_leak"
 ],
 "severity": 1,
 "targets": [
  "framework_or_tooling",
  "enterprise_deployment"
 ],
 "products": [
  "Cursor",
  "Model Context Protocol"
 ],
 "incident_date": "2025-06-24",
 "incident_date_precision": "day",
 "first_reported": "2026-01-30",
 "sources": [
  {
   "url": "https://atlas.mitre.org/studies/AML.CS0045",
   "publisher": "MITRE ATLAS",
   "title": "Data Exfiltration via an MCP Server used by Cursor",
   "published_at": "2026-01-30"
  },
  {
   "url": "https://backslash.security/blog/simulating-a-vulnerable-mcp-server-for-context-poisoning",
   "publisher": "backslash.security",
   "title": "Simulating a Context-Poisoning Vulnerable MCP Server: A POC",
   "published_at": "2026-01-30"
  }
 ],
 "source_count": 2,
 "confidence": 0.95,
 "status": "published",
 "origin": "seed_atlas",
 "tags": [
  "mcp",
  "prompt_injection",
  "cursor",
  "data_leak"
 ],
 "created_at": "2026-10-18T06:55:08.239Z",
 "updated_at": "2026-10-18T06:55:08.239Z"
}